Synopsis The medical evacuation (MEDEVAC) flight departed Wollaston Lake, Saskatchewan, at 2325central standard time en route to La Ronge. After take-off, the aircraft turned about 70degrees to the left, descended, and struck the frozen surface of Wollaston Lake. The pilot and the patient suffered serious injuries; the other two occupants suffered minor injuries. The aircraft was destroyed. The Board determined that, after take-off, the left propeller was likely on its start locks, which, as the airspeed increased, allowed the propeller to overspeed. The pilot was unable to resolve the situation in time to prevent the aircraft from striking the surface of WollastonLake. Contributing to the severity of the patient's injuries were the inadequate restraint provided by the stretcher and its restraining strap, the lack of standards regarding stretchers used in aircraft, and the lack of standards as to the operation of MEDEVAC flights. 1.0 Factual Information 1.1 History of the Flight The aircraft was observed to climb at an unusually shallow angle after take-off, and, when efforts by company personnel to reach the pilot by radio were unsuccessful, a ground search was commenced. The aircraft was found about five minutes after the accident, located on the ice- and snow-covered surface of Wollaston Lake, about 0.75 nautical miles (nm) from the departure end of the runway, and about 1.3 nm from the point of commencement of the take-off roll(3). The pilot and the patient suffered serious injuries. The other two occupants sustained minor injuries. The accident occurred during the hours of darkness at latitude 586.98'N, longitude 10310.79'W, at an elevation of 1,300 feet above sea level (asl). The temperature was about-25C. 1.2 Injuries to Persons 1.3 Damage to Aircraft 1.4 Other Damage 1.5 Personnel Information 1.6 Aircraft Information The aircraft's maintenance records indicate that the aircraft was certified and maintained in accordance with existing regulations and approved procedures. 1.7 Meteorological Information 1.8 Witness Reports The pilot reported that the aircraft's handling and performance was normal during the take-off roll and that he switched on the two combustion heaters immediately after lift-off. Immediately after switching on the heaters, the pilot perceived that the aircraft's engines were no longer running in synchronization, and he believed that there was an engine problem and a possible loss of engine power. There was insufficient yaw effect to indicate which engine was affected, or to indicate the nature of the problem. The pilot reportedly confirmed by reference to the aircraft's instruments that the aircraft was maintaining the runway heading of 340, a pitch attitude of 9 nose-up, and an airspeed of 115 knots. He then confirmed that the engine and propeller controls were fully forward, re-checked the aircraft attitude, and then felt the aircraft's impact with the surface of Wollaston Lake. The pilot estimated that the aircraft was airborne for about four seconds. Witnesses on the ground observed that the aircraft lifted off the runway after a ground run somewhat longer than previous ground runs of that aircraft. No abnormal sounds were noted during the ground roll. After lift-off, the aircraft's angle of climb was observed to be shallower than normal. The aircraft reportedly attained an altitude of about 100 feet above ground level (agl), turned to the left, and disappeared from the view of observers. One of the passengers in the aircraft reported that, just before the impact with the lake, sparks, an arcing sound, and an electrical smell came from the cockpit area in front of thepilot. 1.9 Wreckage and Impact Information The left wing, complete with engine, was found in front of the main wreckage and had been completely severed from the fuselage. The main spar of this wing broke at the fuselage, and both forward and rear wing-to-fuselage attachment points failed in overload. The wing was also severed at the mid-point between the engine and the wing tip, with only the aileron cable connecting the two sections. Damage to the left nacelle indicated that the engine assembly had been pushed upward and rearward at impact. The rear attachment point of the right wing had also failed in overload, and the outer portion of this wing was bent downward approximately 20 degrees, thereby rupturing the fuel tanks. Both propellers had become detached from their respective engines; however, the left propeller had sustained more severe damage. In addition to having more severely bent and twisted blades, the left propeller broke apart in the spider area, thereby allowing one blade to become completely detached from the hub. The nose cone, radome, and main battery were torn from the forward fuselage as the nose structure was forced upward and to the right, thereby crushing the cockpit area, predominantly on the right side. All occupied seats remained in their respective seat rails. Several electrical panels and buses were broken and detached from their mounts. Electrical sparks and arcing sounds and smells may emanate from electrical equipment if it is damaged while under electrical load. The nose baggage door was found about 150 feet from the point of initial impact, and along the centre line of the wreckage trail. The overall damage patterns, the latch status, and the position of the door along the wreckage trail indicate that the door was closed and locked atimpact. At impact, the aircraft was configured with the gear up, flaps up, and all throttle, mixture, and propeller controls fully forward. The fuel selectors were found in the main tank positions while the fuel shut-offs were in the open position. All flight control surfaces were accounted for and control continuity was established. A detailed examination of the airframe revealed no evidence of pre-impact structural failure. 1.10 Fire Both forward and rear combustion heaters were examined and no evidence of fire or pre-impact failure was found. 1.11 Tests and Examination The right engine was not test run, because of the damage it had sustained. The engine was torn down, and all damage found was attributable to overload, impact forces, or heat from the post-crash fire. No pre-existing mechanical faults were revealed which would have contributed to a loss in engine power. The propeller governors were tested, disassembled, and inspected, and both units were found to be functional and operating within accepted parameters. Both propellers were sent to the manufacturer for further investigation. Examination of the left propeller revealed that both start lock stop pins were broken, and that the related area of the high pitch stop sleeve was damaged. The damage signatures and captured blade angles indicate that the engine was rotating and producing high power on impact. The nature of the damage indicated that the start locks were probably engaged and that the left propeller pitch was at its start lock position at the time of impact. This is not normal, and it was not determined why the start lock could have been engaged at the time of impact. The right propeller did not sustain damage to the start locks. The blades showed a distinct power-on twist, although not to such an extent as the left propeller blades. Early separation of the propeller from the engine may have limited the power-on twist signatures of the right propeller blades. On the ground, when the engine is idling, the propeller is on the low pitch stop of 13.2. During shutdown, as the propeller rpm decreases below 800, spring pressure overcomes centrifugal force and the start locks engage. The propeller blades then are free to coarsen by the action of the propeller feathering springs, from the low pitch stop of 13.2 to a point of between 17.2 and 20.2, at which point the start locks prevent further movement. After start up, the blades are driven towards the low pitch stop as oil pressure in the cylinder increases, thereby releasing pressure on the start locks. When the propeller rotation rises above 800rpm, centrifugal force overcomes the spring pressure, extends the start lock pins, and moves them to an outward or retracted position. When the pins are retracted, the propeller blades are free to move throughout their full range of travel. The purpose of the start locks is to prevent the propeller from moving to the feather position after the engine is shut down, so as to reduce drag on the starter during engine start. An engaged start lock pin or pins would allow for blade angle travel between the low pitch stop position of 13.2 and the start lock position (17.2 to 20.2). A propeller with a blade angle at the start lock position will overspeed at certain combinations of engine power and aircraft speed. The cockpit section of the aircraft, complete with the intact instrument panels, was shipped to the TSB Engineering Branch for further examination. Examination did not reveal any evidence of failure that might have resulted in electrical arcing prior to the impact. The engine and flight instruments were examined in an attempt to determine their indications at the time of impact; however, no useful information was found. The vacuum-driven attitude indicator was refrigerated for two days at a temperature of -25C; it was then placed on a test bench and vacuum was applied to it. After a three-minute run-up, the indicator stabilized and performed to normal specifications. The background reference horizon and the white lubber line mark on the surrounding material and the adjustable aircraft symbol (which had not been adjusted after the occurrence) all coincided to indicate zero degrees of pitch. 1.12 Performance Issues According to the aircraft's Pilot's Operating Handbook (POH), the aircraft's maximum rate of climb under the ambient conditions, with both engines operating, would have been 1,370 feet per minute (fpm). With one engine operating and the propeller of the other engine feathered, the maximum rate of climb would be about 410 fpm. According to the aircraft manufacturer, with one engine operating and the propeller of the other engine windmilling, the maximum rate of climb would be under 200 fpm, and if the aircraft were allowed to turn in the direction of the failed engine, the rate of climb would be nil. Maximum climb performance is available only at the aircraft's best rate of climb airspeed. The performance section of the POH lists the best single-engine rate of climb airspeed under the occurrence conditions as 94 knots. The emergency procedures section of the POH indicates that in the event of an engine failure during climb, the pilot should maintain 97knots. Deviation from the aircraft's best rate of climb airspeed will reduce climb performance. According to the aircraft manufacturer's flight test data, at the accident aircraft's take-off weight, and at a pitch angle of 2.53, a constant airspeed of 115 knots will result in level flight, regardless of engine power. At pitch angles of more than 2.53, at 115 knots, the aircraft will climb, and at pitch angles of less than 2.53, at 115 knots, the aircraft will descend. Calculations show that the length of the aircraft's flight path was about 6,000 feet, and at an average speed of 115 knots, the flight would have taken about 30 seconds. If the aircraft began turning as it passed over the departure end of the runway, then the average angle of bank required for the aircraft to reach the crash site would have been about 19. An overspeeding propeller may produce reduced thrust, even though the engine may be producing power, if the blade angle is so fine that the propeller blade is not operating at a positive angle of attack. In that condition, the engine power would drive the propeller to a high rotational speed where the propeller tips exceed the speed of sound, and sonic drag absorbs some of the engine's power. The overspeeding propeller produces less drag than a windmilling propeller, and may produce thrust or drag, depending on propeller rpm and the aircraft's airspeed. The propeller manufacturer estimates that, in this case, if the propeller was on its start locks, it would have produced net thrust at all airspeeds up to and including180 knots. 1.13 Aerodrome Information 1.14 Flight Recorders 1.15 Aircraft Equipment Air Navigation Order Series II, No. 2, provides that a passenger may be carried in a stretcher on an aircraft, provided the stretcher installation system is approved by Transport Canada. An aircraft's stretcher installation system usually comprises a rack or other structure which attaches to hard points in the aircraft cabin and whose upper surface accommodates and secures a stretcher. An approved stretcher installation system may incorporate passenger restraint belts, but they are not mandatory. Transport Canada guidelines indicate that stretchers, aside from their installation systems, are not considered part of the aircraft and do not require approval prior to use in aircraft. Road stretchers are considered acceptable. The operator did not have approval to use the aircraft seats as a stretcher rack or stretcher installation system. 1.16 Training Amendment No. 7 to the syllabus provides that: Captains with more than 50 hours on ME[multi-engine] aircraft with a current PPC need only complete the Recurrent Training. The operator's records indicate that the pilot completed the initial training as required by the training syllabus. The pilot's previous employer operated the Piper PA-31 and PA-34 aircraft types and had authority from Transport Canada to group those two aircraft types for the purpose of pilot proficiency checks (PPC). The effect of such grouping is that a pilot may complete a PPC on each aircraft type in alternate years and retain PPC certification on both types, provided that the required recurrent ground training is completed. The pilot completed a PPC on the PA-31 type in July 1994. He completed a PPC on the PA-34 type in August 1995, and flew the PA-31 type at Eagle Air Services. A pilot's PPC remains current for one year from the end of the month in which the check is completed. When the pilot took up his employment with Eagle Air Services, he completed the required ground training and in each area of study completed the training to initial training standards, but he did not complete another PPC. Transport Canada's policy is that, after a change of the pilot's employment, a PPC which is renewed by an operator's grouping authority is only considered current if the new employer has an equivalent grouping authority. At the time of the occurrence, Eagle Air Services operated the PA-31 and not the PA-34, and did not have authority to group them. The operator reported that the regional Transport Canada, Air Carrier section was consulted at the time the pilot was hired, and that the operator was verbally advised that the pilot's existing PPC would remain valid despite his change of employment, provided the required company training was completed. Grouping authority is reportedly issued on request for most multi-engine aircraft types with maximum allowable take-off weights under 7,000pounds. The pilot's instrument rating was renewed in August 1995 and was valid until 01September1997. Aircraft systems do not allow a propeller overspeed at take-off power to be readily demonstrated or simulated during training. Training on a flight simulator is the only effective way to train for a propeller overspeed emergency at take-off. A simulator is not available for the PA-31 aircraft type. The cockpit indication of a propeller overspeed is a higher-than-normal propeller rpm. The engine manifold pressure and the other engine indications may match those of a normal engine. The differential rudder pressure required to maintain directional control will probably be less than that required in the event of an engine failure, a condition which is commonly simulated during training. There is no direct cockpit warning of a propeller with start locks engaged. Normal pre-take-off procedures require that a propeller check be performed to ensure that the propeller blades will move towards the feather position. The pilot reportedly completed this check during his aircraft run-up checks. The check can be accomplished within the range of propeller blade angle movement which is allowed by a stuck start lock. 1.17 Pilot Workload 1.18 Survival Aspects No pre-flight passenger briefing was provided by the pilot. The pilot believed that the passengers and nurse had flown with the operator before, that the nurse was well-versed on patient restraint issues, and that she would handle patient restraint tasks. In fact, the nurse had only arrived at Wollaston the previous day. When the pilot asked if passengers were restrained, he was told that they were. Air Navigation Order VII, No. 3 requires air operators to provide pre-flight passenger briefings to passengers. The captain of an aircraft has the ultimate authority and responsibility for passenger security. The MEDEVAC was being conducted because of the patient's problems with her pregnancy. The pilot and the nurse were reluctant to restrain the patient over the abdomen, fearing that injury to the fetus might result from doing so. The nurse placed one restraining strap over the patient's lower torso. On impact, the patient's stretcher remained attached to the aircraft's seats, but the patient was thrown from the stretcher and struck the aircraft's interior and interior furnishings. She suffered a serious spinal injury. There was no injury to the fetus. 1.19 MEDEVAC Standards The occurrence flight was arranged by the Wollaston nursing station, an agency of HealthCanada. Most Canadian aviation operation is regulated by the Government of Canada. There are no federal regulations requiring specific training for flight nurses accompanying patients on MEDEVAC flights.